Byzantine Culture

History · Article

The Siege of Constantinople, 717–718

The Umayyad Caliphate's greatest attempt on Constantinople. Leo III, the Bulgar intervention, Greek fire and a winter that destroyed an army — and why the failure mattered for Europe.

Between August 717 and August 718 the Umayyad Caliphate committed its largest army and its largest fleet to the capture of Constantinople. The attempt failed completely. The army that withdrew after twelve months had lost most of its strength to starvation, cold and disease; the fleet was destroyed by fire, storm and desertion. Contemporary Arabic and Byzantine numbers are unusable, but the qualitative agreement across sources in three languages is that the losses were catastrophic and that the caliphate never tried again.

This was the second of the two great Arab sieges — the first, conducted in stages in the 670s, is described in the article on the Heraclian dynasty — and it is the one that decided the question. For the following two centuries the Byzantine–Arab frontier ran through the Taurus mountains and moved by raids and counter-raids rather than by conquest.

The Setting

The caliphate in 715 was at the height of its expansion. Its armies had reached the Indus and the Atlantic and had crossed into Spain in 711; the Byzantine Empire was the one organised state on its borders that had refused to fall. Caliph Sulayman, who succeeded in 715, made the capture of Constantinople the central project of his reign and put his brother Maslama ibn Abd al-Malik in command.

Byzantium, meanwhile, had spent twenty years in political chaos. Between the deposition of Justinian II in 695 and 717 there were seven emperors, most removed by mutiny or coup. The empire was, by any reasonable assessment, ripe for the taking.

Leo III

The man who took the throne in the spring of 717, weeks before the Arab army arrived, was Leo, strategos of the Anatolikon theme — a Syrian by origin, a fluent Arabic speaker, and an experienced frontier commander who had served on diplomatic missions in the Caucasus.

The chronicler Theophanes gives an account of the manoeuvring by which Leo dealt with his predecessor Theodosius III and simultaneously outwitted Maslama, persuading the Arab commander that he would come to terms and inducing him to burn or bypass supplies. Theophanes is writing a century later, is hostile to Leo for his later iconoclasm, and yet still credits him with the defence — which makes the substance of the story more, not less, credible.

Leo entered Constantinople on 25 March 717 and had five months to prepare. He repaired the walls, stocked the granaries, and cleared the city of unnecessary mouths — a measure he enforced by requiring every household to hold three years of provisions or leave.

The Siege

Maslama crossed the Hellespont in July and invested the land walls in August 717. The fleet, under a commander named Sulayman, arrived in September with what the sources describe as eighteen hundred vessels of all types.

Three things went wrong for the attackers within the first weeks.

The walls held. The Theodosian land walls were the most formidable fortification in the world: a triple line of moat, outer wall and inner wall with ninety-six towers. Maslama’s engineers made no impression on them, and the Arab army settled into a blockade.

The fleet was burned. An attempt to force the Golden Horn in September ran into the Byzantine siphon-ships. Greek fire destroyed a squadron at the mouth of the inlet; the rest of the fleet withdrew to anchorages out of range and was never able to complete the blockade by sea. Grain kept reaching the city.

The winter was exceptional. The winter of 717–718 was one of the coldest on record in the region; Theophanes says snow lay for a hundred days. The besieging army, camped in the open in Thrace with insufficient shelter and supply, began to starve. Reports of the eating of pack animals, and worse, appear in both Greek and Arabic accounts.

The Turning Points of 718

In spring two large Egyptian and African squadrons arrived with fresh supplies and troops. They anchored on the Asian side to avoid the fire ships. Their crews were largely Egyptian Christians, and a substantial number deserted to the Byzantines, bringing intelligence on the fleet’s dispositions. Leo attacked with his fire ships at the right moment and destroyed or captured a large part of the relief squadrons.

An overland relief column from Syria was ambushed and destroyed in Bithynia.

Then the Bulgars intervened. Leo had negotiated with Tervel, the Bulgar khan, whose forces attacked the Arab camp from the north. Theophanes claims twenty-two thousand Arab dead in the engagement — again, a number to be treated as an expression of scale rather than a count. The Bulgar attack came at the point when the besiegers were least able to absorb it, and it is the most straightforwardly decisive event of the siege after the failure of the fleet. The relationship between the two powers would soon revert to a long hostility, described in the Byzantine–Bulgarian wars.

Caliph Sulayman had died in 717; his successor Umar II, faced with an army dissolving in Thrace, ordered the withdrawal. Maslama raised the siege on 15 August 718, a year to the day, by some reckonings, after it began. The retreating fleet was hit by a storm in the Aegean and, according to the Greek sources, by an eruption in the volcanic islands; very few ships reached Syria.

Consequences

For the caliphate, the failure ended the strategy of destroying the Byzantine state by a single decisive blow. Umayyad expansion continued in the west and east, but the northern frontier settled into a pattern of annual raiding that neither side expected to be decisive. The financial and manpower losses contributed to the strains that brought down the Umayyads in 750.

For Byzantium, survival gave Leo III unassailable prestige and the political capital to found a dynasty and to reorganise the state: the Ecloga law code of 741, the reform of the themes, and — more divisively — the campaign against religious images that began around 726.

For Europe, the siege is often paired with the Frankish victory at Poitiers in 732 as one of two checks on Arab expansion. The comparison flatters Poitiers. The force Charles Martel defeated was a raiding expedition; the force that broke in front of the Theodosian walls was the main strategic effort of the greatest empire in the world.

The Sources and Their Problems

The Byzantine account comes principally from Theophanes the Confessor and the patriarch Nikephoros, both writing in the late eighth or early ninth century and both hostile to the iconoclast Leo. The Arabic tradition is preserved in al-Tabari and later compilers, drawing on earlier material now lost. The Syriac chronicles offer a third, partly independent line. The three agree on the outline — a year-long siege, a burned fleet, a terrible winter, a Bulgar attack, a shattering withdrawal — and disagree on almost every number and date within it.

There is also a persistent legendary layer, particularly around the figure of Abu Ayyub al-Ansari, a companion of the Prophet said to have died during the first siege of the 670s and whose tomb outside the walls, rediscovered in 1453, became the Eyüp Sultan mosque — the most important Islamic shrine in Istanbul. The Ottoman coronation ceremony of girding the sword took place there, a deliberate link between the sieges that failed and the conquest that succeeded.

The Numbers Problem

Every account of this siege quotes figures, and none of them can be trusted.

Theophanes gives the Arab fleet as 1,800 vessels and the army as 80,000. Arabic sources give comparable or larger numbers. The figure of 120,000 for the besieging force appears in later compilations, and the claim that only five ships returned to Syria is repeated throughout the tradition.

These are not counts. Medieval writers used large round numbers as expressions of magnitude, and both sides had reasons to inflate: the Byzantines to magnify the deliverance, the Arabic tradition to account for a defeat by attributing it to overwhelming commitment.

What can be said with more confidence is qualitative and comes from the consistency of independent traditions. Greek, Arabic and Syriac sources — which do not depend on one another — agree that the force was the largest the caliphate had assembled, that the fleet was destroyed in stages by fire, desertion and storm, that the winter was exceptionally severe, that the besiegers starved, and that the losses were bad enough to end the strategy permanently. That convergence is worth more than any of the figures.

Why the Caliphate Never Tried Again

The failure of 717–718 closed a strategic option, and it is worth being clear about why.

The fleet could not do its job. A siege of Constantinople requires closing the Bosphorus and the Marmara to keep grain out. Greek fire made approaching the Golden Horn suicidal, and a fleet that must anchor out of range cannot blockade.

The land approach is too long. An army besieging the city sits at the end of a supply line running through Anatolia, which the Byzantines could raid, and in Thrace, where the Bulgars could reach it.

The walls were not reducible. Pre-gunpowder siege engineering could not bring down the Theodosian system, as the walls article explains. It took bombards, seven centuries later.

The political cost. The losses contributed materially to the strains that brought down the Umayyad dynasty in 750. The Abbasids who replaced them moved the capital to Baghdad, reoriented toward the east, and treated the Byzantine frontier as a place for annual prestige raids rather than conquest.

The result was a frontier in the Taurus that moved very little for two hundred years, and a relationship of regulated hostility described in Byzantium and the Islamic world.

The Comparison with Poitiers

Pairing 718 with Charles Martel’s victory at Poitiers in 732 is a commonplace of popular history, and it does Poitiers too much credit.

The force Charles defeated was a raiding expedition operating far from its base in Spain. The force that broke in front of Constantinople was the main strategic effort of the largest empire on earth, commanded by the caliph’s brother, supported by the imperial fleet, and intended to end the Byzantine state.

Gibbon’s celebrated passage imagining the Qur’an taught at Oxford had Poitiers in mind. If either battle deserves that kind of counterfactual, it is this one.

Continue exploring